ACTA ICHTHYOLOGICA ET PISCATORIA (2013) 43 (2): 151–161 DOI: 10.3750/AIP2013.43.2.08

CHALLENGES AND OPPORTUNITIES OF LOCAL FISHERIES MANAGEMENT: PIKEPERCH, SANDER LUCIOPERCA (ACTINOPTERYGII: PERCIFORMES: PERCIDAE), IN PÄRNU BAY, NORTHERN GULF OF RIGA, BALTIC SEA

Bärbel MÜLLER-KARULIS 1, 2* , Timo ARULA 3, Maija BALODE 1, Kerli LAUR 3, and Evald OJAVEER 3

1 Latvian Institute of Aquatic Ecology, 8 Daugavgrivas, LV-1007 Riga, Latvia 2 Baltic Nest Institute, Stockholm Marine Centre, Stockholm University, Stockholm, Sweden (present address) 3 Estonian Marine Institute of the University of Tartu, 14 Mäealuse, EE-12618 Tallinn, Estonia

Müller-Karulis B., Arula T., Balode M., Laur K., Ojaveer E. 2013. Challenges and opportunities of local fisheries management: pikeperch, Sander lucioperca (Actinopterygii: Perciformes: Percidae), in Pärnu Bay, northern Gulf of Riga, Baltic Sea . Acta Ichthyol. Piscat. 43 (2): 151–161 . Background. A local stock of pikeperch, Sander lucioperca (Linnaeus, 1758 ), is a valuable fishery resource in Pärnu Bay, northern Gulf of Riga. Due to the late maturity of pikeperch in the bay, the stock is vulnerable to over - exploitation, whereas recruitment is highly dependent on climate factors. Because of its high market price, fish - ing pressure on the stock increased considerably in the 1990s, resulting in stock depression. To assist fishery management in Pärnu Bay we have simulated the effect of several scenarios (such as: fishing mortality, climate, and young-of-the-year food supply) on the pikeperch stock, catches, and revenues generated by the fishery . Materials and methods. A simulation system consisting of an age-structured population model, a recruitment com - ponent, and a forcing module of different climate, food supply to young-of-the-year, and fishing mortality scenarios was used to estimate the equilibrium stock size and catches under varying environmental conditions and exploitation strategies. Economic impacts of these scenarios were assessed based on the first selling price of pikeperch . Results. Under present climate and food supply for pikeperch young-of-the-year, the Pärnu Bay pikeperch stock is very sensitive to catches of immature fish. Warmer future climate conditions are likely to be beneficial for the stock, but also prey abundance for young-of-the-year influences potential stock sizes and catches. Compared to targeting its prey species, herring, which has a lower commercial value at current first selling prices, pikeperch acts as a “biomeliorator”, roughly doubling fisheries revenue in the bay . Conclusion. Flexible adaptive management methods should be used to estimate the yearly allowable pikeperch catch, taking into account food supply to young-of-the-year and climate conditions influencing recruitment . Keywords: Pärnu Bay, pikeperch, temperature, year class abundance, exploitation, fishing mortality, management

INTRODUCTION After Estonia regained independence in 1991, new A local stock of pikeperch, Sander lucioperca export markets for pikeperch opened and privatization of (Linnaeus, 1758 ), adapted to brackish water conditions, is state farms offered local fishermen cheap access to fish - a valuable fishery resource in Pärnu Bay (Vetemaa et al. ing gear. With Estonia’s low average wages in the early 2000, 2006), a shallow bay in the Estonian exclusive eco - 1990s and high first-buyer prices for pikeperch, the aver - nomic zone in the Northern Gulf of Riga (Fig. 1). Despite age catch of pikeperch needed to provide the equivalent of fishing regulations, which were established in Estonia as an average monthly gross wage ranged from 43 kg in early as in the 1920s, this stock has been heavily exploit - 1993 to 240 kg in 1999 (Vetemaa et al. 2000, 2006). ed since the early 1930s, when large-scale export of this Consequently, pikeperch catches increased drastically and fish from Estonia started. After low catches in the 1960s its fishing mortality rose to the highest level ever record - new, biologically justified measures to protect pikeperch ed. Young age groups, including immature fish, became were implemented, such as protected areas and the use of an important part of the catches. This was caused by artificial spawning substrates to increase recruitment. Estonia’s weak legal instruments for regulating fisheries Catches then improved and reached 300 t in the early in the 1990s (Vetemaa et al. 2002) and by insufficient 1990s (Erm et al. 2003) . funding for enforcing the legal minimum size of

* Correspondence: Dr Bärbel Müller-Karulis, Stockholms universitets Östersjöcentrum, Stockholms universitet, SE-106 91 Stockholm, Sweden, phone: +46 8 674 7574, fax: +46 8 16 37 18, e-mail: [email protected] . 152 Müller-Karulis et al. pikeperch in catches (Eero 2004). Pikeperch reach full the final bottleneck for recruitment to the stock maturity only at the age of 5 years, with females maturing (Lappalainen et al. 2002). later (4–5 years) than males (3–4 years), which makes the In addition to climatic fluctuations, eutrophication is stock sensitive to harvesting of immature fish (Erm et a major driving force of ecosystem changes in Pärnu Bay al. 2003). Because fishermen in the 1990s to a large extent and the adjacent Gulf of Riga. In the small, shallow and targeted young age groups before the end of their most rapid comparatively isolated Gulf of Riga surrounded by agricul - growth period, the profits of this fishery were much smaller tural land and industrial centres, eutrophication became evi - than would have been in the case of sustainable manage - dent earlier and was more pronounced than in the open Baltic ment. After 1997 landings decreased sharply (Eero 2004) Sea. Eutrophication increased rapidly from the 1960s and the fishery was temporarily closed in 2000–2003. through the 1980s, had a peak in the late 1980s, and Recruitment success determines the internal dynamics decreased slightly in the 1990s (Nehring et al. 1989, of the Pärnu Bay pikeperch stock (Eero 2004). The shal - Berzinsh 1995, Anonymous 1996, 2002). However, low bay (mean depth 5 m) covers about 700 km 2 and its pikeperch is well known to tolerate or even benefit from salinity ranges from PSU 1 through 5 (Kotta et al. 2009 eutrophication in freshwater systems (Garcia et al. 2006, and references therein). The climate is rather continental, Kangur et al. 2007) and Baltic coastal waters (Lehtonen et with ice covering the bay in winter and the water temper - al. 1996, Lappalainen et al. 2002, Sandström and Karås 2002). ature rising to 24ºC in summer (Arula et al. 2012). Therefore eutrophication most likely poses no direct Pikeperch spawns on sandy bottoms in May and June threat to pikeperch in Pärnu Bay. (Erm et al. 2003). The larvae are planktivorous, first con - Being the largest predator in the Pärnu Bay ecosystem, suming nauplii and young stages of Eurytemora affinis pikeperch converts the biomass of commercially unexploited and Acartia bifilosa (Copepoda). At the end of July fish—gobies ( Pomatoschistus spp.); bleak, Alburnus albur - young-of-the-year (YOY) pikeperch become piscivorous, nus (L., 1758); sticklebacks, Gasterosteus aculeatus L., at first preying chiefly on goby ( Pomatoschistus spp.) lar - 1758; roach, Rutilus rutilus (L.; 1758)—and species of mod - vae as well as Neomysis sp. and Corophium sp., and later erate value as herring, Clupea harengus L., 1758; mainly on juvenile gobies (Erm 1981a). and smelt, Osmerus eperlanus (L., 1758), into biomass of Strong year classes of pikeperch are produced when high commercial value and acts thus as a “biomeliorator”. spawning coincides with warm water and calm weather Therefore, sustainable management of pikeperch (Erm 1981a, b). Further, abundant food, especially goby resources is of high importance for the coastal population larvae and juvenile gobies, when the early juveniles switch around Pärnu Bay. To identify potential management to piscivory improves the growth and consequently sur - strategies for the Pärnu Bay pikeperch population we have vival during the following winter (Lappalainen et al. 2000). developed scenario simulations that address different The wintering conditions for the 0-group pikeperch, exploitation rates, weak or strict enforcement of the legal including winter severity, ice cover duration, abundance size limit, varying prey availability to young-of-the-year, of predators, and fishing intensity in Pärnu Bay, constitute and the effect of climate change on the stock dynamics.

Fig. 1. Location of Pärnu Bay (crosshatched area) in the northern part of the Gulf of Riga, Baltic Sea Pikeperch management in Pärnu Bay, northern Gulf of Riga 153

MATERIALS AND METHODS the Pärnu Bay pikeperch stock (Eero 2004) as calibration To estimate stock sizes and catches of Pärnu Bay data, the final model explained 63.5% of the data deviance 2 pikeperch in equilibrium with different management (R adj = 0.48). All four variables contributed notably to the strategies, exploitation rates, as well as climate and prey explanatory power of the model and were significant at abundance scenarios, we have designed a simulation sys - 0.02 < P ≤ 0.10. Compared to models using only three tem with an age-structured population model for the explanatory variables, each variable improved the explained pikeperch stock, driven by fishing mortality ( F) and envi - deviance by 15.1 percentage points (goby abundance) ronmentally dependent recruitment, as core component. through 27 percentage points (summer temperature). The population model was forced by fishing mortality in The output of the stock-recruitment component was combination with artificial time series capturing expected corrected for the bias introduced by the log transformation climate change and different levels of food supply to (Sprugel 1983) of the recruitment success data. Bias cor - young-of-the-year pikeperch. rection was based on the standard error of the model esti - Pikeperch population model. The population model fol - mates and the approximate degrees of freedom of the lows a standard Virtual Population Analysis (VPA) model GAM model. Further, we have added noise to the output with a Baranov catch equation (Haddon 2011) and of the recruitment model to ensure that average and stan - describes the population dynamics of Pärnu Bay dard deviation of the simulated recruitment corresponded pikeperch using 10 age classes as in Eero (2004). The closely to the mean and standard deviation of the avail - number of fish in age class i at the beginning of year j +1 , able calibration data. The noise component was normally j+1 xNi , is calculated as : distributed with a mean of zero and standard deviation j+1 = j j j Ni Ni 1 exp( Mi 1 Fi 1) [1] corresponding to the average prediction standard error j where M i 1 denotes the natural mortality of age group i estimated by the original GAM stock-recruitment model. j experienced during the previous year, while Fi 1 describes Population and recruitment component coupling. the fishing mortality incurred. Age class 10 survivors Spawning stock biomass (SSB) links recruitment and were returned to age class 10 at the end of each year to pikeperch stock dynamics. Spawning stock biomass SSB j close the population model. at the beginning of year j, is calculated as 10 Integration in time then gives the annual catch weight j j j j j SSB = mi wi Ni [4] in age class i,Ci , as j i=0 j j j j j j Fi mi Ci = Ni wi ()1 ex p( Fi M i ) [2] where is maturity at age i during year j. Both maturity F j + M j i i and weight at age were assumed constant with values j where wi is the weight at age i during year j. In our simu - taken from Eero (2004). lations, we used a constant average weight (Eero 2004) Dynamic simulations. Dynamic simulations of the for each age-class of pikeperch. pikeperch population model were implemented using Stock-recruitment component. We used a generalized ExtendSim (www.extendsim.com) software. The popula - additive model (Hastie and Tibshirani 1986, Wood 2006) tion model was initialized by the observed 1980 pikeperch implemented in the R software package (www.R-proj - stock (Eero 2004) and then run to equilibrium, i.e., until ect.org) to fit the recruitment success of Pärnu Bay stock size and catches showed no further temporal trend, pikeperch to a Ricker-type stock recruitment relation forced by constant fishing mortality in combination with modifiejd by environmental factors: fluctuating climate and prey abundance time-series. The N1 j 1 j 1 j 1 j 1 last 100 years of each simulation run were then used to log j 1 = a + b SSB + c log( G ) + s(T ) + s(I )[3] SSB derive average stock characteristics (total biomass, SSB, j where N1 is the number of one-year olds at the beginning number of recruits) and the magnitude of catches, togeth - j of year j, and factors indexed by j – 1 refer to conditions er with their 5% and 95% percentiles to indicate the mag - during the previous year. Gj – 1 is a trawl index for juve - nitude of stock and catch fluctuations. The spin-up period nile and larval gobies (from here on: gobies) in Pärnu Bay was set arbitrarily to 2000 years to ensure that the stock proportional to the median of goby abundance determined was at equilibrium with the climatic forcing and applied by weekly Hensen trawls above the thermocline during fishing mortality. May–July (Ojaveer et al. 2011, Arula et al. 2012). Tj – 1 Management scenarios. The fishing mortality experi - refers to the average May–July air temperature in Pärnu enced by the Pärnu Bay pikeperch population during the during the year the recruits where spawned, while Ij – 1is time-period covered by VPA data fluctuated between 0.5 the ice duration in Pärnu Bay during their first winter; a, and 2.25 (average 0.89) for mature fish (ages 5–10+), and b, and c are model coefficients, while s() denotes non- between 0 and 0.75 (average 0.08) for the partially mature parametric spline functions. These parameters were ages 3 and 4 (Eero 2004). Comparing the time period selected over a number of other proxies (NAO, winter air before (1960–1989) and after (1993–1998) the breakup of temperature, summer water temperature, salinity, Secchi the Soviet Union (Table 1), fishing mortality increased depth, zooplankton- and mysid abundance) for climatic 2.5-fold for mature fish and the fishing mortality of partial - and feeding conditions in Pärnu Bay to achieve a simple ly mature fish rose from zero to about one tenth (age 3) and model with large R2 and low AIC (Akaike 1974). Using one third (age 4) of the fishing mortality experienced by output from a 1960–1998 virtual population analysis of 5-year olds. 154 Müller-Karulis et al.

Based on these fishing patterns, we constructed two Ministry of the Environment (Fisheries Division), contrasting management scenarios: a) assuming a strict Estonian Ministry of Agriculture, fishermen’s organiza - enforcement of the legal size limit for pikeperch (44 cm) tions, nature conservation bodies, and a group of marine and consequently no catch of immature fish (ages 3–4) scientists in both Estonia and Latvia suggested that and b) significant catch of undersized, immature exploitation of the Pärnu Bay pikeperch stock should pikeperch. For scenario a), selectivity, i.e., the ratio of gradually be adjusted so that mortalities corresponded to fishing mortality for each age group to total fishing mor - a sustainable optimum management of the stock, corre - tality was kept at average values observed in 1960–1989. sponding to its maximum sustainable yield (MSY). The In scenario b) the fishing mortality of ages 3 and 4 was set economic implications of managing Pärnu Bay pikeperch to 10% and 33% of the fishing mortality experienced by at maximum sustainable yield were therefore estimated for 5-year old fish, corresponding roughly to the pattern each fishing mortality setting, prey abundance to YOY and observed in 1993–1998. For older fish, selectivity corre - climate scenario by calculating the revenue generated at sponded to the 1960–1989 average, giving both manage - first sale of pikeperch, based on the first price of €5.0 · kg –1 ment scenarios equal fishing mortality for mature fish. reported by the stakeholders (€ = euro). Further, we com - Climate and prey abundance scenarios. Climate sce - pared this estimate to the potential revenue from fishing narios were based on a) a baseline climate, which cap - herring, which is the major pikeperch prey. The potential tured the statistical characteristics of the observed climate revenue from fishing the annual herring consumption of the used to calibrate the stock-recruitment model for Pärnu Bay pikeperch (Table 2), and b) a future climate scenario, Table 1 where temperature and ice cover duration were shifted to Fishing mortalities experienced by the Pärnu Bay mimic the effects of assumed global warming. pikeperch ( Sander lucioperca ) stock before- To construct forcing time series with sufficient length (1960–1989) and after (1993–1998) to analyse equilibrium conditions between pikeperch the breakup of the Soviet Union, stock and the applied fishing pressure, we first checked calculated from Eero (2004) the input time series—temperature, ice cover duration, Fishing mortality and goby abundance—for autocorrelation. Because all Age autocorrelation functions were small, forcing time series 1960–1989 1993–1998 were generated by a normally distributed random variable 1 0.00 ± 0.00 0.00 ± 0.00 preserving the mean and standard deviation of the input 2 0.00 ± 0.00 0.00 ± 0.00 data. For goby abundance and ice cover transformed data 3 0.00 ± 0.00 0.24 ± 0.15 were used to ensure that forcing functions mimicked the 4 0.01 ± 0.01 0.75 ± 0.36 distribution of the original variables. According to bias-corrected results of regional climate 5 0.47 ± 0.23 2.04 ± 0.79 models the predicted air temperature increase based on 6 0.81 ± 0.39 1.98 ± 0.65 the Intergovernmental Panel on Climate Change (IPCC) 7 0.80 ± 0.30 1.63 ± 0.45 A2 emission scenario in the Gulf of Riga is about 3.5°C 8 0.87 ± 0.38 1.86 ± 0.11 until 2070–2100. In this case at least the central areas of 9 0.68 ± 0.00 1.70 ± 0.50 the basin will be ice free in winter (Sennikovs and Bethers unpublished *). Because the magnitude of expected cli - 10+ 0.68 ± 0.00 1.70 ± 0.50 mate change within this time period is larger than the tem - 3–4 0.00 ± 0.00 0.50 ± 0.23 perature and ice cover range covered by the input data for 5–10+ 0.72 ± 0.16 1.82 ± 0.36 the stock recruitment model (Table 2), we have construct - Values are mean ± standard deviation. ed a hypothetical intermediate future climate scenario by ad ding a moderate increase (1.5°C) to the baseline temper - Table 2 ature time series, setting the ice cover duration simultane - BaselineGclimate- and prey ously to the observed minimum (67 days). Temperatures abundance scenario data (1960–2008) in Pärnu Bay in the future climate scenario exceeding the range covered by the recruitment model input (about 25% of the simulat - Goby Summer air Ice cover ed future temperatures) were capped at the observed max - abundance temperature duration [trawl index] [°C] [days] imum (17.1°C). To mimic the effect of prey abundance to YOY Average 54.6 14.7 130 pikeperch on the pikeperch stock and the fisheries, a low Min 3.00 12.4 67 food availability scenario was generated by setting the Max 306 17.1 164 abundance of gobies to its 5% percentile (trawl index = 22). Standard deviation 60.9 0.98 24.1 In a contrasting, high prey scenario the goby abundance Transformation Log none cubic was set to its 95% percentile (trawl index = 240). Fishery management scenarios. Stakeholder consulta - Transformation = transformation to approximate normal dis - tions in meetings with representatives of the Estonian tribution; Goby = Pomatoschistus spp.

* Sennikovs J., Bethers U. 2009. Statistical downscaling method of regional climate model results for hydrological modelling, 18th World IMACS / MODSIM Congress, Cairns, Australia. S Pikeperch management in Pärnu Bay, northern Gulf of Riga 155 pikeperch stock was based on a first price of €0.15 · kg –1 availability, which had a pronounced effect on the simu - for herring, an annual consumption rate of 6.0 times the lated stock during this period, since fishing mortality total stock biomass, similar to Pärnu Bay piscivore con - simultaneously was at its lowest observed level. sumption in Tomczak et al. (2009), and an 80% share of Effect of harvesting immature fish. Under present cli - herring in the pikeperch diet (Ojaveer et al. 1997). mate conditions and current food supply to YOY, the Pärnu Bay pikeperch stock is very sensitive to catches of RESULTS the age groups 3 and 4 at high exploitation rates (Fig. 3). Model performance. The simulation system (equations While at low to moderate fishing mortality ( F5–10 < 0.8) 1–4) implemented in ExtendSim was validated in two both exploitation patterns reach similar SSB and catches, steps, using as reference dataset the VPA estimates of at higher exploitation rate harvesting immature fish Eero (2004), which the author kindly made available as decreases the SSB and has an even more pronounced number of fish in age classes 1–10 covering 1960–1998. effect on equilibrium catches. Even at the high exploita - First, we validated the recruitment component by initializ - tion rates for mature fish observed after the break-up of ing the stock to its 1960 VPA estimate and then forcing the the Soviet Union ( F5–10 = 1.82) the stock could still have model with the VPA estimates of SSB and the observed sustained a catch of 138 t annually, had there been no temperature, ice cover and goby abundance data used in catch of immature fish, compared to a simulated 38 t with calibrating the stock-recruitment model (Fig. 2, left). Next, such catches. we fully coupled recruitment and population model com - The simulated SSB and catch curves further indicate ponents by using modelled SSB as input to the recruit - that during the Soviet collective fishery management peri - ment component, simulating changes in the stock in od exploitation rates were mostly at equilibrium with 1960–1995 forced now only by the VPA estimates of fish - stock dynamics. VPA estimated SSB and catches mostly ing mortality and the observed environmental conditions fall within the 5%–95% confidence interval of their sim - for recruitment (Fig. 2, right). Gaps in the environmental ulated equilibrium values. Only during 1968–1971 did forcing were replaced by averages in the coupled model catches significantly exceed the equilibrium range, fish - validation, while the corresponding years were omitted in ing down the strong year-classes spawned in the early the recruitment component validation and are consequent - 1960s which lead to a subsequent drop in SSB (Fig. 2, see ly not shown in Fig. 2 (right). Confidence intervals were also Eero 2004). generated as the 5% and 95% percentiles of 1000 simulat - Post-soviet exploitation rates mostly exceeded equi - ed stock trajectories. librium catches, even assuming a strict enforcement of the The recruitment component of the model gives a fair legal size limit (Fig. 3, top right). Targeting also immature description of observed recruitment (Fig. 2, left). The fish, catches greatly exceeded their equilibrium values model reflects the high recruitment observed in the 1960s (Fig. 3, bottom right), causing a rapid stock decline. and around 1990 and corresponds closely to the interme - Pikeperch stock dependence on food supply to young- diate recruitment levels observed during the 1980s. of-the-year. Setting the prey abundance to YOY However, the model overestimates the low reproduction pikeperch to bottom and top 5% of observed values had of the stock observed in the late 1970s and while observed a clear effect on equilibrium SSB and catches (Fig. 4). At SSB declined in the late 1970s and the early 1980s, the low goby abundance SSB and catches declined, while high overestimated recruitment caused a pronounced increase prey abundance led to a pronounced rise in sustainable in the stock. During these years goby abundance in Pärnu harvest and corresponding stock size. Simultaneously the Bay was low. This indicates that the stock-recruitment prediction confidence interval became larger, suggesting model does not fully capture the impact of poor prey that at large stock size equilibrium SSB and annual catch

1600 1200 3 0

1 1400 ] 1000 1

e 1200 g a 800 s

t 1000 i ] t u [ r c B

e 800 600 S r S f

o 600 r 400 n e 400 m u 200 N 200

0 0 1960 1970 1980 1990 2000 1960 1970 1980 1990 2000

Fig. 2. Number of recruits at age 1 (left) and spawning stock biomass (right) simulated by the pikeperch (Sander lucioperca ) population model (lines: model results (continuous) with 5% and 95% percentiles (dashed), dots: VPA estimates (Eero 2004, Eero, personal communication) 156 Müller-Karulis et al. will fluctuate more. In the model, the increased fluctua - Revenue generated at maximum catch. Under our first tions at high stock size are produced by the exponential selling price assumptions, the annual revenue generated functions applied in the catch equation (Equation 2) and in by harvesting pikeperch at maximum equilibrium catch the stock-recruitment function (Equation 3). rates varies from €614 000 per year under unfavourable Impact of expected climate change. Expected climate YOY feeding conditions to €2 474 000 per year in change has an even larger positive effect on equilibrium a potential warmer climate (Table 3). Significant catches SSB than high goby abundance. Our simulations indicate of immature fish decrease the annual revenue by approx - that the stock might sustain catches of close to 500 t annu - imately 14% from €814 000 per year to €704 000 per ally at a SSB of about 800 t (Fig. 5). year, assuming the fishing mortality can be maintained at

1600 450

1400 400

1200 350 ]

t 300 1000 [ ] h t [ c 250 t B 800 a S C

S 200 600 150 400 100 200 50 0 0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 F5–10 F5–10 1600 450 1400 400

1200 350

] 300 t [ ] 1000 t [ h

c 250 B t

S 800 a S C 200 600 150 400 100 200 50 0 0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 F5–10 F5–10 Fig. 3. Spawning stock biomass (left) and catch (right) in equilibrium with different levels of fishing mortality, assum - ing no (top) or significant catch (bottom) of immature, undersized pikepearch ( Sander lucioperca ); Lines show average simulated SSB and catch (continuous), with 5% and 95% percentiles (dashed); Markers denote observed SSB and catches (Eero 2004) with no (circles) or significant (triangles) catch of immature fish; F5–10 = average fishing mortality age 5–10 Table 3 Fishing mortality, annual catch, equilibrium total stock biomass and revenue generated at maximum catch of pikeperch (Sander lucioperca ) for simulated scenarios, compared to potential revenue generated from harvesting equivalent herring prey

Fishing Catch Total stock Annual herring Pikeperch Equivalent Scenario –1 (SSB in brackets) consumption revenue herring revenue mortality [t·year ] [t] [t·year–1] [k€ ·year–1] [k€ ·year–1] PCPANIP 0.9 163 582 (401) 2794 814 419 PCPASIP 0.54 140 677 (497) 3249 701 487 PCLPANIP 0.72 123 491 (353) 2357 614 353 PCHPANIP > 2.16 338 912 (536) 4378 1690 657 FCPPANIP > 2.16 495 1354 (787) 6500 2474 975 k€ = thousands of euro; PCPANIP = Present climate and prey abundance to YOY, no catch of immature pikeperch; PCPASIP = Present climate and prey abundance to YOY, significant catch of immature fish; PCLPANIP =Present climate, low prey abundance to YOY, no catch of immature pikeperch; PCHPANIP =Present climate, high prey abundance to YOY, no catch of immature pikeperch; FCPPANIP = “Future” climate, present prey abundance to YOY, no catch of immature pikeperch. SSB = spawning stock biomass. Pikeperch management in Pärnu Bay, northern Gulf of Riga 157 its low equilibrium value. Compared to targeting herring, was found in coastal areas of the Southern Baltic (Gröger et fishing pikeperch was economically beneficial in all sce - al. 2007), most likely caused by cannibalism at high num - narios and approximately doubled the revenue generated. ber of recruits (Winkler 1989). Pikeperch cannibalism is Targeting pikeperch was particularly advantageous in sce - also well known from lakes, especially at high juvenile narios with good pikeperch reproduction conditions (no density (Frankiewicz et al. 1999, Lappalainen et al. 2006), catch of immature pikeperch, high prey abundance to but has so far not been documented in Pärnu Bay. Intra- YOY, future climate), because high recruitment increases specific competition between juveniles in strong year- the yield per stock biomass and consequently generates classes has also been suggested to make pikeperch recruit - higher yield for each unit of herring consumed by ment density-dependent (Lappalainen et al. 2009). pikeperch. Incorporating environmental variables can enhance the performance of stock-recruitment models (Fiksen and DISCUSSION Slotte 2002, Marjomäki 2004, Keyl and Wolff 2008) and The stock-recruitment model constructed for the GAM models are a flexible tool for incorporating non-linear Pärnu Bay pikeperch population confirms the importance environmental effects (Daskalov 1999, Chen et al. 2005, of water temperature, winter severity, and abundance of Megrey et al. 2005, Keyl and Wolff 2008). Even though goby for recruitment found by earlier analysis of long- the pikeperch stock-recruitment model presented here term monitoring data (Erm 1981a, b, Lappalainen et performs well, with most data points within the 95% con - al. 2000). In addition, it predicts a decline of recruitment fidence limit of the model (Fig. 2), overestimated repro - success at high SSB (coefficient b < 0 in equation 3). duction during consecutive years in the late 1970s gave In the modelled scenarios, this effect stabilizes the equi - a pronounced deviation between observed and simulated librium SSB and catch curves at large fishing pressure, SSB in the early 1980s, resembling the ordinary error because the smaller SSB is partly offset by a larger num - amplification effect in stepwise models (Håkanson ber of recruits per spawner, creating the slowly tailing 1999, 2003). catch- curves in scenarios describing beneficial recruitment In all modelled scenarios, the variability in predicted conditions (expected climate change, high goby abun - catches is largest at high fishing mortality, good overall dance). A similar weak negative relation between recruitment conditions (high goby abundance, climate pikeperch spawning stock biomass and recruitment success change), and stock sizes, when the stock removal through

1200 450 400 1000 350 800 300 ] t ] [ t

[ 250 h c B

600 t S a S 200 C 400 150 100 200 50 0 0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 F5–10 F5–10 2500 700

600 2000 500 ] t ] [ t 1500

[ 400 h c B t S a S

C 300 1000 200 500 100

0 0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 F5–10 F5–10 Fig. 4. Selected population characteristics of the Pärnu Bay pikepearch ( Sander lucioperca ): Spawning stock biomass (left) and catch (right) in equilibrium with different levels of fishing mortality, assuming low (top) or high (bot - tom) food supply expressed as goby abundance; Lines show average simulated SSB and catch (continuous), with 5% and 95% percentiles (dashed). Markers denote observed SSB and catches (Eero 2004) with no (circles) and significant (triangles) catch of immature fish; F5–10 = average fishing mortality age 5–10 158 Müller-Karulis et al. catches is roughly balanced by a gain in recruitment suc - implying a change in weight at age. Faster growth in turn cess. Similar density-dependent effects have been shown causes earlier maturity in pikeperch populations adapted to to contribute to fluctuations in “Norwegian” cod warmer conditions (Lappalainen et al. 2003). Therefore the (Bjørnstad et al. 1999) and marine crab populations Pärnu Bay pikeperch stock might be less sensitive to the (Higgins et al. 1997). Moreover, in our scenarios this effect catch of small fish in a future warmer climate. is amplified by high fishing pressure, which immediately Goby abundance in Pärnu Bay has declined since the removes strong year-classes from the stock when the fish 1960s (Laur et al. 2009) and is mainly related to low water reach the legal size limit. Theoretical analysis of age-struc - transparency (Laur et al. unpublished *), which in turn tured models has shown variability to increase with depends on wind-induced resuspension, ship traffic, and exploitation rate, reproductive success, and maturation age peat processing in the shallow bay (Paavel et al. 2011). of the population (Andrews et al. 2004). This effect has also Many fish larvae feeding on zooplankton depend on been found in empirical studies of North Sea cod (Andrews vision in their search for prey (Guthrie 1986) and therefore et al. 2004) and marine crabs (Higgins et al. 1997). increased turbidity of the water can reduce feeding and Our scenarios indicate a positive effect of the expect - reaction distance. On the other hand, suspended particles ed climate change towards higher temperatures on the interfere with detection by predators (Utne-Palm 2002). pikeperch stock. However, the average simulated stock Pikeperch has a “spying” feeding behaviour and is size at F5–10 = 1 is about 1000 t, fluctuating between 300 favoured by low water transparency (Erm 1981b and refer - and 1800 t (5% and 95% percentiles). The maximum ences therein). Thus, the interactions between goby abun - stock size used to construct the stock-recruitment model dance, turbidity and prey abundance for YOY pikeperch was 720 t, implying that a large part of the stock sizes are complex, which is also expressed by the relatively low simulated in scenarios beneficial to pikeperch reproduc - contribution of goby abundance to the explained deviance tion (high goby abundance, expected climate change) are in the stock-recruitment model and by the underestimation outside the range for which the stock-recruitment relation of recruitment during low goby abundance. has been calibrated. Lappalainen et al. (2009) have sug - Our estimates of the pikeperch fishery revenue in gested that density dependent effects on recruitment comparison to targeting their main prey species, herring, might increase in a warmer climate, making the effect of clearly indicate that pikeperch acts as a “biomeliorator” in climate change on pikeperch populations more complex. Pärnu Bay. However, these estimates are sensitive to the Further, our model does not take into account whether first selling price of both pikeperch and herring. Pärnu Bay can provide enough food for a doubled or According to the stakeholder consultations fishermen and tripled pikeperch population. A recent food web model fishing companies perceive a high market demand for (Tomczak et al. 2009) indicates that in Pärnu Bay pisci - pikeperch and expect high prices also in the future. vores, especially the quantitatively dominating perch, Therefore stakeholders preferred to introduce a sustain - exert the strongest top–down control on planktivores able optimum management scheme by adjusting the within the five coastal ecosystems studied. This suggests pikeperch fishing mortality to maximize the catch from that food competition will play a role if there is a large the stock. To achieve this goal, the following measures increase in piscivorous fish populations. In a warmer future were suggested: 1) the mortality of immature pikeperch climate, also some assumptions underlying our model should be reduced considerably. Combined methods might no longer hold. Higher temperatures can be expected should be applied to limit catches (temporal/spatial clo - to increase the growth rates of many fish species in the sures, limitations by gear types or by professional/non- Baltic (MacKenzie et al. 2007 and reference therein), professional fishermen) 2) allowable annual catches 4500 1200 4000 1000 3500 3000 800 ] t ] [ t 2500 [ h 600 c B t S 2000 a S C 1500 400 1000 200 500 0 0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 F5–10 F5–10 Fig. 5. Selected population characteristics of the Pärnu Bay pikepearch ( Sander lucioperca ): Spawning stock biomass (left) and catch (right) in equilibrium with different levels of fishing mortality at “future” climate conditions; Lines show average simulated SSB and catch (continuous), with 5% and 95% percentiles (dashed); Markers denote observed SSB and catches (Eero 2004) with no (circles) and significant (triangles) catch of immature fish; F5–10 = average fishing mortality age 5–10

* Laur K., Ojaveer H., Simm M., Klais R. Multidecadal dynamics of goby larvae Pomatoschistus spp. in response to environmental variability in a shallow temperate bay. Submitted to Estuarine, Coastal and Shelf Science. Pikeperch management in Pärnu Bay, northern Gulf of Riga 159 should be calculated taking into account the natural factors Anonymous 1996. Third periodic assessment of the state of the determining the strength of pikeperch year classes, i.e. Baltic Sea 1989 – 1993. Background Document. HELCOM temperature conditions during spawning and the first win - Anonymous 2002. Fourth periodic assessment of the state of the tering period. The clear emphasis on reducing the catch of marine environment in the Baltic Sea area, 1994–1998. undersized fish is based on the past experience in manag - HELCOM. ing the stock, when after high fishing pressure on imma - Arula T., Kotta J., Lankov A., Simm M., Põlme S. 2012. Diet ture fish not even a complete closure of the fishery was composition and feeding activity of larval spring-spawning sufficient to recover the stock. Our scenarios indicate that herring: Importance of environmental variability. Journal of the pikeperch stock can sustain catches of 140 t · year –1 at Sea Research 68 : 33–40. DOI: 10.1016/j.seares.2011.12.003 significant harvest of immature fish, which is only Berzinsh V. 1995. Hydrological regime. Pp. 7–31. In : Ojaveer E. a decrease by 14% compared to the maximum catch at (ed.) Ecosystem of the Gulf of Riga between 1920 and 1990. harvesting only mature fish. However, without restricting Vol. 5. Estonian Academy Publishers, Tallinn, Estonia. the catch of immature individuals the equilibrium fishing Bjørnstad O.N., Fromentin J.-M., Stenseth N.C., Gjøsæter J. mortality is very low (0.54 compared to 0.9 at enforcing 1999. Cycles and trends in cod populations. Proceedings of the legal size limit, see Table 3), making the stock very the National Academy of Sciences of the United States of sensitive to overfishing. The models constructed quanti - America 96 (9): 5066–5071. DOI: 10.1073/pnas.96.9.5066 fied the relation between environmental conditions and Chen Q.X., Chan K.S., Lekve K., Torstensen E., Gjøsæter J., sustainable catch and it was decided that the contacts Ottersen G., Stenseth N.C. 2005. Population dynamics of between the participating organizations for gradual cod Gadus morhua in the North Sea region: Biological den - approximation of the management of the stock to the sus - sity-dependent and climatic density-independent effects. tainable optimum stage should be continued in the future. Marine Ecology-Progress Series 302 : 219–232. DOI: 10.3354/meps302219 CONCLUSIONS Daskalov G. 1999. Relating fish recruitment to stock biomass Pikeperch acts as a “biomeliorator” for the Pärnu Bay and physical environment in the Black Sea using generalized fishery, roughly doubling the revenues generated from fishing additive models. Fisheries Research 41 (1): 1–23. DOI: the amount of herring needed to sustain the pikeperch popu - 10.1016/S0165-7836(99)00006-5 lation. Our analysis indicates that the mortality of immature Eero M. 2004. Consequences of management of pikeperch pikeperch must be reduced considerably to achieve catches at (Stizostedion lucioperca L.) stock in Pärnu Bay (Baltic Sea) the maximum sustainable yield of the stock. Since recruit - under two different economic regimes, 1960–1999. ment depends in a complex way on climate conditions and Fisheries Research 68 (1–3): 1–7. DOI: 10.1016/ prey abundance, flexible adaptive methods should be used to j.fishres.2004.03.002 calculate the yearly allowable catch. For the implementation Erm V. 1981a. Koha. [Pikeperch.] Valgus, Tallinn, Estonia. [In of an optimum strategy for the sustainable management of the Estonian.] pikeperch stock, collaboration between the scientists, Erm V. 1981b. Populâcionnye parametry i ocenka stepeni fishermen’s organizations, management bodies and min - èkspluatacii zapasov sudaka v Pârnuskoj buhte. [Population istry administrators should be continuous. parameters and estimation of the exploitation rate of pikeperch in Pärnu Bay.] Rybohozâjstviennye issledovaniâ ACKNOWLEDGEMENTS v bassejne Baltijskogo morâ, Riga 16 : 46–63. This work was funded by the SPICOSA project (EU Erm V., Vaino V., Saat T. 2003. Pikeperch, Sander lucioper - FP 6) and received additional support from the Estonian ca —Fishes of Estonia. Pp. 296–306. In : Ojaveer E., Pihu E., Ministry of Education and Research (grant Saat T. (eds.) Fishes of Estonia. Estonian Academy, Tallinn, SF0180005s10), the Estonian Science Foundation (grant Estonia. 8747) and the European Regional Fund (grant Fiksen Ø., Slotte A. 2002. Stock-environment recruitment mod - 3.2.0802.11-0029). The authors would like to thank the els for Norwegian spring spawning herring ( Clupea haren - entire SPICOSA project team, in particular Ragnar gus ). Canadian Journal of Fisheries and Aquatic Sciences 59 Elmgren, Tom Hopkins, and Denis Bailly for their sug - (2): 211–217. DOI: 10.1139/f02-002 gestions to the text, and Margit Eero for providing the Frankiewicz P., Dabrowski K., Martyniak A., Zalewski M. pikeperch data. We also would like to thank both anony - 1999. Cannibalism as a regulatory force of pikeperch, mous reviewers for their careful work. Stizostedion lucioperca (L.), population dynamics in the lowland Sulejow reservoir (Central Poland). Hydrobiologia REFERENCES 408–409 (0): 47–55. DOI: 10.1023/A:1017001803791 Akaike H. 1974. A new look at the statistical model identifica - Garcia X.-F., Diekmann M., Brämick U., Lemcke R., tion. IEE Transactions on Automatic Control 19 (6): Mehner T. 2006. Correlations between type-indicator fish 716 –723. DOI: 10.1109/TAC.1974.1100705 species and lake productivity in German lowland lakes. Andrews J.M., Blythe S.P., Gurney W.S.C. 2004. Stability Journal of Fish Biology 68 (4): 1144–1157. DOI: analysis of a continuous age-structured model with specific 10.1111/j.0022-1112.2006.01009.x reference to North Sea cod. Journal of Biological Systems Gröger J.P., Winkler H., Rountree R.A. 2007. Population 12 (3): 249–260. DOI: 10.1142/S0218339004001191 dynamics of pikeperch ( Sander lucioperca ) and its linkage 160 Müller-Karulis et al.

to fishery driven and climatic influences in a southern Baltic of environmental and economic factors. Archive of Fishery lagoon of the Darss-Zingst Bodden chain. Fisheries and Marine Research 49 (3): 199–212. Research 84 (2): 189–201. DOI: 10.1016/j.fishres. Laur K., Arula T., Ojaveer H. 2009. Long-term dynamics of 2006.10.018 early life stages of gobies ( Pomatoschistus spp.) in Pärnu Guthrie D.M. 1986. Role of vision in fish behaviour. Pp. Bay (Gulf of Riga). Pp. 231–231. In : Abstract book: 7th 75–113. In : Pitcher T.J. (ed.) The behaviour of teleost fish - Baltic Sea Science Congress 2009; 17–21 August 2009, es. Croom Helm, Beckenham, Kent, UK . DOI: 10.1007/978- Tallinn University of Technology, Estonia. 1-4684-8261-4_4 Lehtonen H., Hansson S., Winkler H. 1996. Biology and Haddon M. 2011. Modelling and quantitative methods in fish - exploitation of pikeperch, Stizostedion lucioperca (L), in the eries. 2nd edn. CRC Press, Boca Raton, FL, USA. Baltic Sea area. Annales Zoologici Fennici 33 (3–4): Håkanson L. 1999. Error propagations in step-by-step predic - 525–535. tions: examples for environmental management using MacKenzie B.R., Gislason H., Möllmann C., Köster F.W. regression models for lake ecosystems. Environmental 2007. Impact of 21st century climate change on the Baltic Modelling and Software 14 (1): 49–58. DOI: Sea fish community and fisheries. Global Change Biology 13 10.1016/S1364-8152(98)00022-X (7): 1348–1367. DOI: 10.1111/j.1365-2486.2007.01369.x Håkanson L. 2003. Propagation and analysis of uncertainty in Marjomäki T.J. 2004. Analysis of the spawning stock— ecosystem models. Pp. 139–167. In : Canham C.D., Cole J.J., Recruitment relationship of vendace ( Coregonus albula Lauenroth W.K. (eds.) Models in ecosystem science. (L.)) with evaluation of alternative models, additional vari - Princeton University Press, Princeton, New Jersey, USA. ables, biases and errors. Ecology of Freshwater Fish 13 (1): Hastie T., Tibshirani R. 1986. Generalized additive models. 46–60. DOI: 10.1111/j.0906-6691.2004.00041.x Statistical Science 1 (3): 297–318. DOI: 10.1214/ Megrey B.A., Lee Y. -W., Macklin S.A. 2005. Comparative ss/1177013604 analysis of statistical tools to identify recruitment— Higgins K., Hastings A., Sarvela J.N., Botsford L.W. 1997. Environment relationships and forecast recruitment strength. Stochastic dynamics and deterministic skeletons: Population ICES Journal of Marine Science 62 (7): 1256–1269. DOI: behavior of Dungeness crab. Science 276 (5317): 10.1016/j.icesjms.2005.05.018 1431–1435. DOI: 10.1126/science.276.5317.1431 Nehring D., Schulz S., Rechlin O. 1989. Eutrophication and Kangur K., Park Y. -S., Kangur A., Kangur P., Lek S. 2007. fishery resources in the Baltic. Journal du Conseil Patterning long-term changes of fish community in large International pour l’Exploration de la Mer 190 : 198–205. shallow Lake Peipsi. Ecological Modelling 203 (1–2): Ojaveer E., Arula T., Lankov A., Shpilev H. 2011. Impact of 34–44. DOI: 10.1016/j.ecolmodel.2006.03.039 environmental deviations on the larval and year-class abun - Keyl F., Wolff M. 2008. Environmental variability and fisheries: dances in the spring spawning herring ( Clupea harengus what can models do? Reviews in Fish Biology and Fisheries membras L.) of the Gulf of Riga (Baltic Sea) in 1947–2004. 18 (3): 273–299. DOI: 10.1007/s11160-007-9075-5 Fisheries Research 107 (1–3): 159–168. DOI: Kotta J., Kotta I., Simm M., Põllupüü M. 2009. Separate and 10.1016/j.fishres.2010.11.001 interactive effects of eutrophication and climate variables on Ojaveer H., Lankov A., Lumberg A., Turovski A. 1997. the ecosystem elements of the Gulf of Riga. Estuarine, Forage fishes in the brackish Gulf of Riga ecosystem (Baltic Coastal and Shelf Science 84 (4): 509–518. DOI: Sea). Pp. 293–309. In : Forage fishes in marine ecosystems. 10.1016/j.ecss.2009.07.014 Proceedings of the International Symposium on the Role of Lappalainen J., Dörner H., Wysujack K. 2003. Reproduction Forage Fishes in Marine Ecosystems. Anchorage, Alaska, biology of pikeperch ( Sander lucioperca (L.)) —a review. USA, November 13–16, 1996. Alaska Sea Grant College Ecology of Freshwater Fish 12 (2): 95–106. Program, University of Alaska Fairbanks. DOI: 10.1034/j.1600-0633.2003.00005.x Paavel B., Arst H., Metsamaa L., Toming K., Reinart A. Lappalainen J., Erm V., Kjellman J., Lehtonen H. 2000. 2011. Optical investigations of CDOM-rich coastal waters Size-dependent winter mortality of age-0 pikeperch in Pärnu Bay. Estonian Journal of Earth Sciences 60 (2): (Stizostedion lucioperca ) in Pärnu Bay, the Baltic Sea. 102–112. DOI: 10.3176/earth.2011.2.04 Canadian Journal of Fisheries and Aquatic Sciences 57 (2): Sandström A., Karås P. 2002. Effects of eutrophication on 451–458. DOI: 10.1139/f99-270 young-of-the-year freshwater fish communities in coastal Lappalainen J., Milardi M., Nyberg K., Venäläinen A. 2009. areas of the Baltic. Environmental Biology of Fishes 63 (1): Effects of water temperature on year-class strengths and 89–101. DOI: 10.1023/A:1013828304074 growth patterns of pikeperch ( Sander lucioperca (L.)) in the Sprugel D.G. 1983. Correcting for bias in log-transformed allo - brackish Baltic Sea. Aquatic Ecology 43 (1): 181–191. DOI: metric equations. Ecology 64 (1): 209–210. DOI: 10.1007/s10452-007-9150-y 10.2307/1937343 Lappalainen J., Olin M., Vinni M. 2006. Pikeperch cannibal - Tomczak M.T., Müller-Karulis B., Järv L., Kotta J., Martin ism: effects of abundance, size and condition. Annales G., Minde A., Põllumäe A., Razinkovas A., Strake S., Zoologici Fennici 43 (1): 35–44. Bucas M., Blenckner T. 2009. Analysis of trophic networks Lappalainen A., Soderkultalahti P., Wiik T. 2002. Changes in and carbon flows in south-eastern Baltic coastal ecosystems. the commercial fishery for pikeperch ( Stizostedion lucioper - Progress in Oceanography 81 (1–4): 111–131. DOI: ca ) on the Finnish coast from 1980 to 1999—Consequences 10.1016/j.pocean.2009.04.017 Pikeperch management in Pärnu Bay, northern Gulf of Riga 161

Utne-Palm A.C. 2002. Visual feeding of fish in a turbid envi - bility in the Estonian coastal fisheries sector. Marine Policy ronment: Physical and behavioural aspects. Marine and 30 (6): 635–640. DOI: 10.1016/j.marpol.2005.08.001 Freshwater Behaviour and Physiology 35 (1–2): 111–128. Winkler H.M. 1989. The role of predators in fish communities DOI: 10.1080/10236240290025644 in shallow coastal waters of the southeast Baltic. Rapports et Vetemaa M., Eero M., Hannesson R. 2002. The Estonian fish - Proces-verbaux des Réunions. Conseil International pour eries: from the Soviet system to ITQs and quota auctions. l’Exploration de la Mer 190 : 125–132. Marine Policy 26 (2): 95–102. DOI: 10.1016/S0308- Wood S.R. 2006. Generalized additive models. An introduction 597X(01)00040-9 with R. Chapman and Hall/CRC, Texts in Statistical Vetemaa M., Eschbaum R., Aps R., Saat T. 2000. Collapse of Science, Boca Ration, FL, USA. political and economical system as a cause for instability in fisheries sector: an Estonian case. Pp. 1–9 . In : Proceedings of the IIFET 2000 International Conference Microbehaviour and Macroresults; 10–14 July 2000, Oregon State University, Corvallis, OR, USA. Received: 29 October 2012 Vetemaa M., Eschbaum R., Saat. T. 2006. The transition from Accepted: 29 May 2013 the Soviet system to a market economy as a cause of insta - Published electronically: 30 June 2013